排序方式： 共有3条查询结果，搜索用时 171 毫秒
1
1.
We propose in this paper an architecture for near-duplicate video detection based on: (i) index and query signature based structures integrating temporal and perceptual visual features and (ii) a matching framework computing the logical inference between index and query documents. As far as indexing is concerned, instead of concatenating low-level visual features in high-dimensional spaces which results in curse of dimensionality and redundancy issues, we adopt a perceptual symbolic representation based on color and texture concepts. For matching, we propose to instantiate a retrieval model based on logical inference through the coupling of an N-gram sliding window process and theoretically-sound lattice-based structures. The techniques we cover are robust and insensitive to general video editing and/or degradation, making it ideal for re-broadcasted video search. Experiments are carried out on large quantities of video data collected from the TRECVID 02, 03 and 04 collections and real-world video broadcasts recorded from two German TV stations. An empirical comparison over two state-of-the-art dynamic programming techniques is encouraging and demonstrates the advantage and feasibility of our method. 相似文献
2.
Relevance feedback methods generally suffer from topic drift caused by word ambiguities and synonymous uses of words. Topic drift is an important issue in patent information retrieval as people tend to use different expressions describing similar concepts causing low precision and recall at the same time. Furthermore, failing to retrieve relevant patents to an application during the examination process may cause legal problems caused by granting an existing invention. A possible cause of topic drift is utilizing a relevance feedback-based search method. As a way to alleviate the inherent problem, we propose a novel query phrase expansion approach utilizing semantic annotations in Wikipedia pages, trying to enrich queries with phrases disambiguating the original query words. The idea was implemented for patent search where patents are classified into a hierarchy of categories, and the analyses of the experimental results showed not only the positive roles of phrases and words in retrieving additional relevant documents through query expansion but also their contributions to alleviating the query drift problem. More specifically, our query expansion method was compared against relevance-based language model, a state-of-the-art query expansion method, to show its superiority in terms of MAP on all levels of the classification hierarchy. 相似文献
3.
Image-based modeling is an appropriate technique to create 3D models of cultural heritage objects, which starts with the basic task of designing the camera network. This task is, however, quite crucial in practical applications because it needs a thorough planning and a certain level of experience. The optimal camera network is designed when certain accuracy demands are fulfilled with a reasonable effort, namely keeping the number of camera shots at a minimum. In this study, we report on the development of an automated method for designing the optimal camera network for a given cultural heritage building or statue. Starting from a rough point cloud derived from a video image stream, the initial configuration of the camera network is designed, assuming a high-resolution HR state-of-the-art non-metric camera. To improve the image coverage and accuracy, we use a mathematical non-linear optimization with constraints. Furthermore, synthetic images are created to guide the camera operator to the designed images. From the first experimental test, we found that a target accuracy of 10 mm could be maintained although the initial number of more than 300 high-resolution images got reduced to less than 90 for the final, optimized network. 相似文献
1